Tortelli, Daniel MeninMeneguel, Giovani André2022-05-202022-05-202021http://repositorio.uricer.edu.br/handle/35974/430The use of innovative technologies is gaining more and more prominence in everyday systems and utilities. Computer vision algorithms fit into this scenario, enabling its usability in the auto- mation and simplification of diverse activities in the most varied areas. Linked to the structure of an embedded system, this technology can be innovative and useful, bringing more accuracy and speed to repetitive tasks, such as counting small-scale objects. The objective of this work was to build an embedded system that uses computer vision algorithms in order to recognize and perform object counting. The development of system features made use of the Python language. Django and TensorFlow frameworks were used for web development and detection model trai- ning, respectively. The data set, as well as the images used for detection, were captured through the embedded system. The hardware structure of the embedded system were developed using the products accrued from the Raspberry Pi.pt-BRCiência da ComputaçãoVisão computacionalSistemas embarcadosDetecção de objetosTrabalho de Conclusão de Curso